We examined the criteria for sleeping place selection in a social band of Rhinopithecus bieti (black-and-white snubnosed or golden monkeys) living in the mountainous Samage Forest, Baima Snow Mountain Nature Reserve, Yunnan, China. We performed principal component analysis and found that slope aspect, tree height and trunk diameter were likely key variables influencing selection of sleeping places. Sleeping sites were preferentially located in mixed deciduous/conifer forest. The monkeys slept exclusively in evergreen trees, of which 82% were conifers (mostly Picea likiangensis and Tsuga dumosa) and 18% evergreen oaks (Cyclobalanopsis oxyodon and Quercus spp.). Sleeping trees were tall (mean 30.5 m), had high boles (mean 18.4 m), large diameters (mean 62.6 cm) and large crown areas (mean 57.9 m 2 ). A comparative analysis of phytological and architectural features between trees in "sleeping site plots" (n = 18) and trees in "non-sleeping-site plots" (n = 66) revealed that diameter, crown surface area and tree height were significantly (P < 0.01) larger in the former compared with the latter. All investigated roosting sites were situated on steep mountain slopes. Valleys and mountain ridges were avoided. We also detected re-use of roosting sites on several occasions, but not on consecutive nights. It is most likely that a mix of factors (stability of trees, access to food, unit cohesion, monitoring potential) explains the pattern of sleeping site preference, but predation at night seems to be only slightly important. Climate appears to have a profound influence on patterns of sleeping site selection in the monkeys' harsh temperate habitat. This is demonstrated by the monkeys' preference for mixed forest at medium elevations over montane fir forest at high elevations and slopes instead of ridges, with reduced exposure to wind and precipitation inherent in the former. We also emphasize the possibly substantial role that non-environmental factors (the nature of social organization and socio-behavioral strategies) play in determining sleeping site use in R. bieti and other primates.